Binary Neural Networks (BNN) have binarized neuron and connection values so that their accelerators can be realized by extremely efficient hardware. However, there is a significant accuracy gap between BNNs and networks with wider bit-width. Conventional BNNs binarize feature maps by static globally-unified thresholds, which makes the produced bipolar image lose local details. This paper proposes a multi-input activation function to enable adaptive thresholding for binarizing feature maps: (a) At the algorithm level, instead of operating each input pixel independently, adaptive thresholding dynamically changes the threshold according to surrounding pixels of the target pixel. When optimizing weights, adaptive thresholding is equivalent to an accompanied depth-wise convolution between normal convolution and binarization. Accompanied weights in the depth-wise filters are ternarized and optimized end-to-end. (b) At the hardware level, adaptive thresholding is realized through a multi-input activation function, which is compatible with common accelerator architectures. Compact activation hardware with only one extra accumulator is devised. By equipping the proposed method on FPGA, 4.1% accuracy improvement is achieved on the original BNN with only 1.1% extra LUT resource. Compared with State-of-the-art methods, the proposed idea further increases network accuracy by 0.8% on the Cifar-10 dataset and 0.4% on the ImageNet dataset.

In 2014, the above paper entitled ‘Quasi-Linear Support Vector Machine for Nonlinear Classification’ was published by Zhou, et al. [1]. They proposed a quasi-linear kernel function for support vector machine (SVM). However, in this letter, we point out that this proposed kernel function is a part of multiple kernel functions generated by well-known multiple kernel learning which is proposed by Bach, et al. [2] in 2004. Since then, there have been a lot of related papers on multiple kernel learning with several applications [3]. This letter verifies that the main kernel function proposed by Zhou, et al. [1] can be derived using multiple kernel learning algorithms [3]. In the kernel construction, Zhou, et al. [1] used Gaussian kernels, but the multiple kernel learning had already discussed the locality of additive Gaussian kernels or other kernels in the framework [4], [5]. Especially additive Gaussian or other kernels were discussed in tutorial at major international conference ECCV2012 [6]. The authors did not discuss these matters.

Advanced driver-assistance systems (ADAS) generally play an important role to support safe drive by detecting potential risk factors beforehand and informing the driver of them. However, if too many services in ADAS rely on visual-based technologies, the driver becomes increasingly burdened and exhausted especially on their eyes. The drivers should be back out of monitoring tasks other than significantly important ones in order to alleviate the burden of the driver as long as possible. In-vehicle auditory signals to assist the safe drive have been appealing as another approach to altering visual suggestions in recent years. In this paper, we developed an in-vehicle auditory signals evaluation platform in an existing driving simulator. In addition, using in-vehicle auditory signals, we have demonstrated that our developed platform has highlighted the possibility to partially switch from only visual-based tasks to mixing with auditory-based ones for alleviating the burden on drivers.

Recent character detectors have been modeled using deep neural networks and have achieved high performance in various tasks, such as text detection in natural scenes and character detection in historical documents. However, existing methods cannot achieve high detection accuracy for wooden slips because of their multi-scale character sizes and aspect ratios, high character density, and close character-to-character distance. In this study, we propose a new U-Net-based character detection and localization framework that learns character regions and boundaries between characters. The proposed method enhances the learning performance of character regions by simultaneously learning the vertical and horizontal boundaries between characters. Furthermore, by adding simple and low-cost post-processing using the learned regions of character boundaries, it is possible to more accurately detect the location of a group of characters in a close neighborhood. In this study, we construct a wooden slip dataset. Experiments demonstrated that the proposed method outperformed existing character detection methods, including state-of-the-art character detection methods for historical documents.

We propose a non-photorealistic rendering method for automatically generating point-light-style (PLS) images from photographic images using peripheral difference filters with different window sizes. The proposed method can express PLS patterns near the edges of photographic images as dots. To verify the effectiveness of the proposed method, experiments were conducted to visually confirm PLS images generated from various photographic images.

For massive multiple-input multiple-output (MIMO) communication systems, simple linear detectors such as zero forcing (ZF) and minimum mean square error (MMSE) can achieve near-optimal detection performance with reduced computational complexity. However, such linear detectors always involve complicated matrix inversion, which will suffer from high computational overhead in the practical implementation. Due to the massive parallel-processing and efficient hardware-implementation nature, the neural network has become a promising approach to signal processing for the future wireless communications. In this paper, we first propose an efficient neural network to calculate the pseudo-inverses for any type of matrices based on the improved Newton's method, termed as the PINN. Through detailed analysis and derivation, the linear massive MIMO detectors are mapped on PINNs, which can take full advantage of the research achievements of neural networks in both algorithms and hardwares. Furthermore, an improved limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) quasi-Newton method is studied as the learning algorithm of PINNs to achieve a better performance/complexity trade-off. Simulation results finally validate the efficiency of the proposed scheme.

In this paper, a circuit based on a field programmable analog array (FPAA) is proposed for three types of chaotic spiking oscillator (CSO). The input/output conversion characteristics of a specific element in the FPAA can be defined by the user. By selecting the proper characteristics, three types of CSO are realized without changing the structure of the circuit itself. Chaotic attractors are observed in a hardware experiment. It is confirmed that the dynamics of the CSOs are consistent with numerical simulations.

We thank Kamata et al. (2023) [1] for their interest in our work [2], and for providing an explanation of the quasi-linear kernel from a viewpoint of multiple kernel learning. In this letter, we first give a summary of the quasi-linear SVM. Then we provide a discussion on the novelty of quasi-linear kernels against multiple kernel learning. Finally, we explain the contributions of our work [2].

Various optical fiber connectors have been developed during the 40 years since optical fiber communications systems were first put into practical use. This paper describes the key technologies for optical connectors and recent technical issues.

We have proposed autonomous network diagnosis platform for operation of future large capacity and virtualized network, including 5G and beyond 5G services. As for the one candidate of information collection and analyzing function blocks in the platform, we proposed novel optical sensing techniques that utilized tapped raw signal data acquired from digital coherent optical receivers. The raw signal data is captured before various digital signal processing for demodulation. Therefore, it contains various waveform deformation and/or noise as it experiences through transmission fibers. In this paper, we examined to detect two possible failures in transmission lines including fiber bending and optical filter shift by analyzing the above-mentioned raw signal data with the help of machine learning. For the purpose, we have implemented Docker container applications in WhiteBox Cassini to acquire real-time raw signal data. We generated CNN model for the detections in off-line processing and used them for real-time detections. We have confirmed successful detection of optical fiber bend and/or optical filter shift in real-time with high accuracy. Also, we evaluated their tolerance against ASE noise and invented novel approach to improve detection accuracy. In addition to that, we succeeded to detect them even in the situation of simultaneous occurrence of those failures.

We present our data-collection and deep neural network (DNN)-training scheme for extracting the optical status from signals received by digital coherent optical receivers in fiber-optic networks. The DNN is trained with unlabeled datasets across multiple administrative network domains by combining federated learning and unsupervised learning. The scheme allows network administrators to train a common DNN-based encoder that extracts optical status in their networks without revealing their private datasets. An early-stage proof of concept was numerically demonstrated by simulation by estimating the optical signal-to-noise ratio and modulation format with 64-GBd 16QAM and quadrature phase-shift keying signals.

This work shows a broadband, high-efficiency power amplifier (PA) monolithic microwave integrated circuit (MMIC) that uses InP-based metal-oxide-semiconductor (MOS) high-electron-mobility transistors (HEMTs) with an extended drain-side access region and broadband conjugate matching topology. Advanced device technologies, namely, double-side-doped structures, MOS gate structures, and asymmetric gate recess, were adopted, and the length of the drain-side access region was optimized to simultaneously obtain high power and efficiency. A common-source PA MMIC based on InP-based MOS-HEMTs was fabricated, and an interstage circuit was designed to maximize the S21 per unit stage in the broadband, resulting in a record-high power-added efficiency and wide bandwidth.

This paper presents a compact fully-differential distributed amplifier using a coupled inductor. Differential distributed amplifiers are widely required in optical communication systems. Most of the distributed amplifiers reported in the past are single-ended or pseudo-differential topologies. In addition, the differential distributed amplifiers require many inductors, which increases the silicon cost. In this study, we use differentially coupled inductors to reduce the chip area to less than half and eliminate the difficulties in layout design. The challenge in using coupled inductors is the capacitive parasitic coupling that degrades the flatness of frequency response. To address this challenge, the odd-mode image parameters of a differential artificial transmission line are derived using a simple loss-less model. Based on the analytical results, we optimize the dimensions of the inductor with the gradient descent algorithm to achieve accurate impedance matching and phase matching. The amplifier was fabricated in 0.18-µm CMOS technology. The core area of the amplifier is 0.27 mm2, which is 57% smaller than the previous work. Besides, we demonstrated a small group delay variation of ±2.7 ps thanks to the optimization. the amplifier successfully performed 30-Gbps NRZ and PAM4 transmissions with superior jitter performance. The proposed technique will promote the high-density integration of differential traveling wave devices.

This paper demonstrates a phase compensation technique using varactors for variable-gain phase shifters (VGPSs). The VGPS consists of an I/Q generator and I/Q variable gain amplifiers (I/Q VGAs). I/Q VGAs based on common-emitter stages are enabled to control the gain by adjusting the collector current of the transistor. However, the phase control performance degenerates because the input capacitance varies with the collector current. The proposed phase compensation technique reduces the variation in the insertion phase of the I/Q VGA by adjusting the voltage of the varactor provided at its input and maintaining the input capacitance constant in any gain state. As a result, the VGPS can provide a low phase and amplitude error under phase control. A Ka-band VGPS with the proposed phase compensation technique, fabricated in a 130-nm SiGe BiCMOS process, demonstrates a 0.73° and 0.06 dB improvement in the RMS phase and amplitude error compared with the case without the compensation technique. The VGPS achieves measured RMS amplitude and phase errors of less than 0.19 dB and 0.75°, respectively, in an amplitude control range of more than 20 dB with a frequency range of 28 to 32 GHz.

This contribution introduces a novel, dielectric waveguide based, permittivity sensor. Next to the fundamental hybrid mode theory, which predicts exceptional wave propagation behavior, a design concept is presented that realizes a pseudo-transmission measurement approach for attenuating feed-side reflections. Furthermore, a transmission line length independent signal processing is introduced, which fosters the robustness and applicability of the sensor concept. Simulation and measurement results that prove the sensor concept and validate the high measurement accuracy, are presented and discussed in detail.

The generation and reduction reaction of magnesium plasma were studied using a cylindrical transverse magnetic-mode applicator in magnetic and electric field modes. By heating Mg powder using the magnetic field mode, plasma was generated with the evaporation of Mg and stably sustained. When the Mg plasma sample was introduced into the reaction zone and exposed to microwave and lamp heating, a reduction reaction of scandium oxide also occurred. The results of this study provide prospects for the development of a larger microwave refining system.

Precision microwave measurement systems require highly stable oscillators with both excellent long-term and short-term stability. Compared to components used in laboratory instruments, dielectric resonator oscillators (DRO) offer low phase noise with greatly reduced mechanical complexity. To further enhance performance, phase-locked loop (PLL) stabilization can be used to eliminate drift and provide precise frequency control. In this work, the design of a low-cost DRO concept is presented and its performance is evaluated through simulations and measurements. An open-loop phase noise of -107.2 dBc/Hz at 10 kHz offset frequency and 12.8 GHz output frequency is demonstrated. Drift and phase noise are reduced by a PLL, so that a very low jitter of under 29.6 fs is achieved over the entire operating bandwidth.

We evaluated the energy efficiency per 1-bit transmission of an optical light source on InP substrate to achieve optical interconnection. A semiconductor optical amplifier (SOA) assisted extended reach EADFB laser (AXEL) was utilized as the optical light source to enhance the energy efficiency compared to the conventional electro-absorption modulator integrated with a DFB laser (EML). The AXEL has frequency bandwidth extendibility for operation of over 100Gbit/s, which is difficult when using a vertical cavity surface emitting laser (VCSEL) without an equalizer. By designing the AXEL for low power consumption, we were able to achieve 64-Gbit/s, 1.0pJ/bit and 128-Gbit/s, 1.5pJ/bit operation at 50°C with the transmitter dispersion and eye closure quaternary of 1.1dB.

The optical phased array (OPA) is an emerging non-mechanical device that enables high-speed beam steering by emitting precisely phase-controlled lightwaves from numerous optical antennas. In practice, however, it is challenging to drive all phase shifters on an OPA in a deterministic manner due to the inevitable fabrication-induced phase errors and crosstalk between the phase shifters. In this work, we fabricate a 16-element silicon photonic non-redundant OPA chip with integrated phase monitors and experimentally demonstrate accurate monitoring of the relative phases of light from each optical antenna. Under the beam steering condition, the optical phase retrieved from the on-chip phase monitors varies linearly with the steering angle, as theoretically expected.

We present an effective system for integrating generative zero-shot classification modules into a YOLO-like dense detector to detect novel objects. Most double-stage-based novel object detection methods are achieved by refining the classification output branch but cannot be applied to a dense detector. Our system utilizes two paths to inject knowledge of novel objects into a dense detector. One involves injecting the class confidence for novel classes from a classifier trained on data synthesized via a dual-step generator. This generator learns a mapping function between two feature spaces, resulting in better classification performance. The second path involves re-training the detector head with feature maps synthesized on different intensity levels. This approach significantly increases the predicted objectness for novel objects, which is a major challenge for a dense detector. We also introduce a stop-and-reload mechanism during re-training for optimizing across head layers to better learn synthesized features. Our method relaxes the constraint on the detector head architecture in the previous method and has markedly enhanced performance on the MSCOCO dataset.